Pre-RA: Antibodies Sort the Risk, Ultrasound Erosions Flag Severity Save

Three ACR 2026 abstracts from the Leeds CCP cohort tested how to triage anti-CCP positive individuals with musculoskeletal symptoms. Serology and a few clinical items carried the prediction, while ultrasound erosions spoke to severity after arthritis began.

Pre-RA is the phase in which autoantibodies and joint symptoms are present without clinical synovitis. Many people in it never develop arthritis, and rheumatology clinics cannot follow all of them. Three talks from the Leeds CCP cohort at ACR Convergence 2026 asked who can return to primary care, who is close to onset and what imaging adds.

Sana Sharrack (Leeds) looked for good prognosis (Abstract 1835). Of 450 CCP-positive at-risk individuals with five years of follow-up, 317 (70%) stayed arthritis-free. At ten years the figure was 73 of 157 (46%). Five baseline features independently predicted non-progression at five years: anti-CCP below 3 times the upper limit of normal, negative RF, normal ESR, one tender joint or fewer and no tenosynovitis on ultrasound. With 3 or more of them, 82% were arthritis-free at five years and 19 of 21 at ten. All 39 patients with all five stayed free at five years. The ten-year groups are small. The proposal is community follow-up for this group.

Laurence Duquenne (Leeds) asked the opposite question: can a score applied at any visit identify the 6 months before onset (Abstract 1836)? The analysis used 2,353 visits from 543 participants, of whom 176 (32%) progressed after a median of 1 year. In the last 6 months before arthritis, RF, morning stiffness, small joint tenderness and ultrasound power Doppler and tenosynovitis all rose. Four markers made the final score: anti-CCP at 3 times the upper limit or higher, positive RF, morning stiffness of 30 minutes or more and at least one tender small joint. With all four present, 46% progressed within 6 months. Against the published Leeds baseline score, the C-index was 0.864 vs 0.856, with no significant difference. Both scores lean on the same serology, so the baseline score can be reused at later visits.

Imaging found its role in the first talk. Andrea Di Matteo (Leeds) followed 176 CCP-positive individuals after they progressed to arthritis (Abstract 1834). Before onset, 55 (31%) had ultrasound bone erosions, mostly at MTP5, and 30% of eroded joints showed no synovitis. Pre-arthritis erosions predicted b/tsDMARD initiation (HR 1.71, 95% CI 1.01-2.88) and difficult-to-treat RA (HR 3.63, 95% CI 1.31-10.05). Radiographic erosions showed a trend (HR 2.30, 95% CI 0.93-5.66). In discussion, Di Matteo noted that anti-CCP did not predict difficult-to-treat disease in this analysis. Erosions did.

The sorting tools are ready for clinic. Serology plus a short list of clinical items identifies a large low-risk group who can leave secondary care, and repeated scoring adds little to the baseline result. Ultrasound erosions answer a different question, how severe the disease will be once it starts. None of these tools has yet been shown to change outcomes when used to choose patients for prevention. The next data to watch are the Leeds decision trees and the first prevention trials that enroll by these risk groups.